Mathematics Internal Assessment

Modeling India’s GDP

November 2022

1
Table of Contents

I) Introduction and Rationale ...........................................................................................................3

1. Aim and Approach ...................................................................................................................3
II) Data plan .....................................................................................................................................4
III) Mathematical Exploration .........................................................................................................5
1. Exponential Modeling ..............................................................................................................7
2. Logistic modeling ..................................................................................................................11
IV) Analysis ...................................................................................................................................17
1. Exponential model .................................................................................................................17
2. Logistic Model .......................................................................................................................18
V) Conclusion ................................................................................................................................18
VI) Evaluation and Extension ........................................................................................................18
VII) Bibliography...........................................................................................................................19
VIII) Appendices ...........................................................................................................................20
1. Appendix A - Chi-square calculations for the Exponential Model........................................20
2. Appendix B - Chi-square calculations for the Logistic Model ..............................................22

2
I) Introduction and Rationale
In my economics class we learnt about the business cycle, a graph depicting the fluctuations in a
nation’s economic activity, which is usually a measure of GDP (Gross Domestic Product).

The real GDP of a country (adjusted to changes in currency fluctuations) showcases the total
value of all goods and services produced in a year.

A country’s GDP is an informative indicator that drives countless decision-making processes, for
instance, national economic policies and interest rates at the macroeconomic level and budgets
and multi-year plans at the microeconomic level. This makes the forecast of GDP a relevant
issue, and one very much applicable in real-life scenarios, which is why I aim to use mathematics
to model and construct equations that fit India’s GDP.

What I am most interested in finding out is the extent to which any model I construct will be able
to predict the future GDP. There are several methods used to calculate GDP, but the expenditure
method taught in my class was: C: Evidence of
significant personal
GDP = C + I + G + (X-M) engagement.

- C = consumer expenditure on goods and services

- I = investment expenditure by financial institutions, as well as households on real-estate
- G = government expenditure
- X = exports
- M = imports

The formula showcases the sheer number of factors that can influence GDP, and even those
factors are influenced by several others. I learnt that even a minute change in one of these factors
will affect the GDP as a whole, which is why I’m curious to see if a model can actually predict
the changes in every single one of these factors, and as a result, GDP as a whole.

1. Aim and Approach

Hence, my aim of the investigation is to establish an ideal equation that models India’s GDP
using calculus. The model will relate GDP with time.

3
 A: Clear aim stated

There are many mathematical models available to represent the GDP of the country, and I will
attempt to apply the one which is most relevant to my data. Then, I will decide the optimum
model for making future predictions using the chi-square test.

II) Data plan

I have collected the data of India’s real GDP from 1960 to 2019 from the World Bank, and
tabulated it below.

Table 1 : India's GDP for the years 1960-2019 (GDP (Constant 2015 US$) - India | Data)

Year GDP/Billion Year GDP/Billion

units units

1960 136.68 1990 465.243

1961 141.445 1991 470.16
1962 145.591 1992 495.936
1963 154.318 1993 519.496
1964 165.819 1994 554.089
1965 161.448 1995 596.059
1966 161.359 1996 641.058
1967 173.987 1997 667.02
1968 179.882 1998 708.271
1969 191.645 1999 770.923
1970 201.529 2000 800.534
1971 204.84 2001 839.152
1972 203.707 2002 871.073
1973 210.42 2003 939.543
1974 212.914 2004 1014
1975 232.395 2005 1094
1976 236.26 2006 1183
1977 253.4 2007 1273
1978 267.876 2008 1312
1979 253.844 2009 1416
4
 1980 270.943 2010 1536
1981 287.216 2011 1616
1982 297.199 2012 1705
1983 318.861 2013 1813
1984 331.044 2014 1948
1985 348.438 2015 2104
1986 365.082 2016 2277
1987 379.559 2017 2432
1988 416.102 2018 2591
1989 440.849 2019 2696

Using these 60 values, I constructed a scatter plot, to observe the relationship between real GDP
and time, shown in Figure 1. B: Appropriate use
of table and graph

Figure 1: Scatter Plot of India's real GDP, based on Table 1.

III) Mathematical Exploration

5
By carefully observing the above graph, it shows that the function is increasing, which can fit
different mathematical models.

To decide which model applies to India’s GDP, I first defined the variables that will be involved
in the model.

Independent variable – time, measured in years (T)

B: Variables are
Dependent variable – real GDP (y) defined

I could observe, looking at the graph that the higher India’s GDP was, the faster India’s GDP
increased. The rate of change of India’s GDP with respect to time, is proportional to India’s GDP
itself. This gave me:

dy
=ky, where k is the constant of proportionality.
dT

This is a separable differential equation which can be solved easily.

dy
Here, dT, the rate of change of y with respect to T, is not a fraction. It is the instantaneous rate of
change, the limit, but to be able to solve this equation I treated these derivatives in Leibniz
notations like they were fractions and the differentials like they were quantities. Using Leibniz’s
notation over Lagrange notation and Newton’s notation was so that it was clear what I was
differentiating with respect to, and allowed me to effectively carry out a dimensional analysis.

A separable differential equation can be solved by separating the variables. So, I moved all the
quantities with y to one side of the equation and t to the other. To do this, I divided both sides by
y, which gave me:

1 dy 1
× =ky×
y dT y

1 dy 1
× =ky×
y dT y

1 dy
× =k
y dT

Then I integrated both sides of the equation:

6
 1 dy
∫ × dT= ∫ k dT
y dT

ln|y|=kt+c1

Ideally there would be another constant on the left side of the equation as well, but the next step
would’ve been to subtract the constant from both sides thereby omitting the need to include it in
the first place.

|y|=e(kt+c1) → log rules C: Personal

engagement is
evident
|y|=ekt ×ec1 →laws of exponents

ec1 is a constant that I defined as a new constant C, so |y|=Cekt , and since the GDP cannot be
negative, y > 0 and |y|= y.

Thus, I have simplified the differential equations solution to be of the form y=Cekt . This is an
exponential equation, suggesting that India’s GDP can be modeled using exponential modeling.

1. Exponential Modeling

Now that I knew the function was in the general exponential equation form of y=Cekt with 𝑦,
with the dependent variable as the real GDP and t, the independent variable as the year, I could
make some conclusions about the nature of the exponential equation by observing the scatter
plot.

1. The exponent x has a positive coefficient – the curve slopes upward, as x→ 0, y→ 0. This
can only occur if the exponent is positive, and thus k > 0.
2. In the graph, as x→ ∞, y → ∞ as well. Therefore, the graph slopes upward and C, the
coefficient of ekt is positive.

To find the constants C and k, I substituted values from the data collected:

(1990, 465.243)

465.243=Ce1990k →1

(1960, 136.68)
7
 2696=Ce2019k →2

465.243
Using equation 1, C = e1990k

465.243
Substituting this into equation 2 gave me 2696 = e2019k
e1990k

2696=465.243e29k

e29k =5.794821201
E: Clear
explanation
ln 5.794821201=29k demonstrating
good
k=0.06058498698 ~0.0606 understanding of
concept
There are 2 different equations that can be used to find C -

465.243 465.243
C= = ≈1.969568532×10-50
e1990k e1990×0.0606

2696 2696
C= 2019k
= 1990×0.0606
≈1.968711212×10-50
e e

The values obtained for C are nearly identical, so the value of C was rounded to C=1.97×10-50 .

Substituting the values for C and k into the function gave me y=(1.97×10-50 )×e0.0606t . When
graphed on the modeling software Desmos (“Desmos | Graphing Calculator”), the equation
produced the following curve:

8
 Figure 2: An exponential model of India's GDP

Without performing any mathematical calculations, I saw that the curve deviates from the data
points in some instances. If the curve deviates from the data points, the accuracy of any
extrapolations becomes questionable. This is why another model must be used.

Quadratic and cubic models were considered, but were eliminated for the reason that they both
have a maximum turning point, after which the modeled curve would inevitably decrease.
Therefore, the model cannot be extrapolated with accuracy. Trigonometric models were
eliminated for this same reason. D: Meaningful
reflection seen
It was then that I was reminded of an article I read in the London School of Economics (LSE)
blog site, about whether infinite economic growth is possible on a finite planet. (“Is Infinite
Economic Growth Possible on a Finite Planet?”)

The world’s resources are limited, and so GDP growth rate will get smaller and smaller,
approaching a maximum imposed by the depleted resources.

It can be said that the environment can’t support a GDP more than M, the maximum, in which
case, the exponential model is not applicable beyond a certain point.

dy
This means that the same equation I used in exponential modeling, =ky is feasible provided y
dT

is well below M. However, as y → M, dy/dT→0.

9
 dy y
I demonstrated this relationship in an algebraic expression, =ky (1- M). For smaller values of y,
dt
y y dy y
is also small, and (1- ) is close to 1, so is high. For larger values of y, is bigger, and (1-
M M dt M
y dy
) is closer to 0 and decreases, thus solving the issue that arose in exponential modeling.
M dt

dy y
To test the equation that has been derived, =ky (1- M), I sketched its graph. I used 2 reference
dt

points, f(t) = 0 and f(t) = M.

If f(t) = 0 as in GDP = 0, as time passes, it will continue to remain 0. More output can only be
made using previously produced output.

If f(t) = M as in GDP = M, as time passes, GDP continues to remain M due to the constraints
imposed by the environment.

For 0 < f(t) < M, if f(t) is close to 0, ky is the only factor that influences the gradient and so,
y
gradient increases in proportion to GDP. If f(t) is closer to M, (1- M) influences GDP more, and

gradient starts to decrease once more. After drawing these conclusions, I sketched a graph:

Figure 3: A sketch of the ideal model of India’s GDP

By looking at the sketch as well as the scatter plot, I deduced that India’s GDP can be modeled
using the equation I derived, which is called the logistic differential equation. Though the scatter
plot only resembled the initial portion of the sketch, logic suggests that the growth of GDP
10
would be more suitable for a logistic model rather than exponential, and thus, I modeled India’s
GDP using logistic modeling.

2. Logistic modeling

To construct a logistic model to graph India’s GDP, I expanded and simplified the derived
dy y
equation =ky (1- M).
dt

y
Multiplying both sides by dt gave me dy=ky (1- M) dt.

y
Then, I divided both sides by y(1- M ).

1
y dy=k dt
y(1- M )

Next, I expanded the left side into 2 fractions using partial fraction expansion.

1 M M
y × =
y (1- M) M y(M-y)

A B M
+ =
y M-y y(M-y)

A B M
( + = ) y(M-y)
y M-y y(M-y)
E: Correct use of
M=A(M-y)+By mathematics

1M+0=AM-Ay+By

1M+0=(A)M+(B-A)y

A=1

B-A=0

B-1=0

11
 B=1

1 1 M
+ =
y M-y y(M-y)

1 1
+ =k dt
y M-y

1 1
∫( + ) dy = ∫ k dt
y M-y

1
I then integrated M-y -

1
∫ du , u=M-y
u

1
∫ du = ln|u|, ∫ u dy =-1
u

1
∫ = -ln |M-y|
M-y
B: Neat
Using this,
presentation
1 1
∫ + = ∫ k dt
y M-y

ln|y| -ln|M-y|+C1 =kt+C2

I wrote C2-C1 as a new constant C3, and removed the modulus signs (as GDP > 0).

ln y -ln M-y =kt+C3

y
ln =kt+C3
M-y

y
=ekt+C3
M-y

y
=ekt ×eC3
M-y

12
eC3 is another constant that I named C.

y
=Cekt
M-y

M-y
=Ce-kt
y

M-y C: Outstanding
y× =y×Ce-kt
y personal
engagement is
M-y=yCe-kt evident

M-y+y=y+yCe-kt

M=y+yCe-kt

M=y(1+Ce-kt )

M
Y=
(1+Ce-kt )

Thus, I have derived the general equation of my logistic model, which has 3 unknowns, M, C
and k, rather than the two M and k in the exponential model.

Therefore, before I could use simultaneous equations, I knew that I had to eliminate one variable,
so I did some research.

I found out that, due to the consequences of climate change damage, negative economic
consequences would be especially large in South and South-East Asia. (Dellink et al.)

D:Evidence of
meaningful
reflection

13
 Figure 4: OECD data on the effects of climate change on GDP

OECD’s prediction for India’s real GDP by 2060 is 42,203,870 million US dollars (“Domestic
Product - GDP Long-Term Forecast - OECD Data”), or 42203.87 billion US dollars at 2010
Purchasing Power Parities (PPP’s).

However, the GDP values used to construct these models were in 2015 PPP’s. PPPs are used in
order to eliminate the influence of inflation and currency fluctuations on GDP, so it is vital that
the year used for the PPP’s be uniform. So, I converted the 42,203,870 million US dollars into
2015 PPP’s. (OECD)

42203.87
×19.235=55605.92601
14.599

Being guardedly optimistic, I took this value to be M, the maximum.

55605.92601
Y=
(1+Ce-kt )

Now, I could use simultaneous equations:

(1990, 465.243)

55605.92601
465.243=
(1+Ce-1990k )
14
(2019, 2696)

55605.92601
2696=
(1+Ce-2019k )

55605.92601=465.243(1+Ce-1990k )

465.243(1+Ce-1990k )
2696=
(1+Ce-2019k )

(1+Ce-1990k ) E: Good
5.79=
(1+Ce-2019k ) knowledge and
understanding are
5.79(1+Ce-2019k )=1+Ce-1990k demonstrated

(5.79+5.79Ce-2019k )=1+Ce-1990k

4.79+5.79Ce-2019k =Ce-1990k

Inputting this equation into Wolfram Alpha (4.79+5.79xe^(−2019y)=x(E^(−1990y)) -

Wolfram|Alpha) with C = x and k = y gave me:

Figure 5: A simplification performed by Wolfram Alpha

55605.92601
465.243= -1990k
479e2019k
(1+ e )
100e29k -579

-1990k
479e2019k
(1+ e )=90.71
100e29k -579

15
Inputting this into a graphic display calculator (CASIO fx-CG50) with x = k gave me k ≈ 0.062.

B: Use of
technology

Figure 6: A screenshot of the Graphic Calculator

479e2019×0.062
=C
100e29×0.62 -579

C≈4.47×1055 C: Evidence of
significant personal
55605.92601 engagaement
Thus, the logistic equation of India’s GDP is Y= -0.062t
1+(4.47×1055 e )

Inputting the equation into Desmos gave me –

16
 Figure 7: A logistic model of India’s GDP

Figure 8: A logistic model of India‘s GDP, with the data points

IV) Analysis
To evaluate the accuracy of the models, and determine which of the models best fit the actual
data values, I made use of the chi-square goodness of fit test.

I defined the null and alternate hypothesis as –

H0: there is no significant difference between the expected and actual data values.

H1: there is a significant difference between the expected and actual data values.

Based on the nature of the data being explored I deemed a significance level of 0.1% acceptable
as an estimate of future GDP is what businesses and other institutions require to predict
quantities such as sales. An exact value is both impossible to estimate and improbable to occur in
real life due to the unstable nature of the economy.

1. Exponential model
The chi-square value for the exponential model is 569.4095826 (Appendix A). In order to accept
or reject the null hypothesis, the critical value must be found. The degrees of freedom is the

17
(number of categories - 1), so 60-1 = 59. With a df of 59, and a significance level of 0.1, the
critical value is 98.324. (National Institute of Standards and Technology).

Since the chi-square value of 569.4095826 > 98.324, the null hypothesis is rejected, and the
alternate hypothesis is accepted. As the difference between the chi-square value and the critical
value was so large, I decided to verify my results using the p-value. The p-value was
1.36975×10-84, supporting the results, and proving that there is a significant difference between
the expected and actual values and it is not due to chance variation.

2. Logistic Model
The chi-square value is 655.0214992. Since 655.0214992 > 98.324, the null hypothesis is
rejected, and the alternate hypothesis is accepted. The p value for the logistic model was
1.878151×10-101, once again rejecting the null hypothesis, and accepting the alternate hypothesis.

V) Conclusion
When constructing the models, it appeared to me that they would be an accurate model of India’s
GDP, as it visually fit the data points. However, when using the chi-squared test, I began to see
that it was not as accurate as I initially thought; in fact, there was nearly no evidence supporting
the null hypothesis.

Although this result was not expected, it showed me the importance of relying on statistical tools
to draw conclusions. Mathematically, my models are correct, so I can only assume that the
models’ inability to predict GDP lies in the unstable nature of GDP itself. It is subject to both
long- and short-term fluctuations, which I was aware of when beginning this investigation. This
may hinder its ability to be predicted by a static model.
D: Evidence of
VI) Evaluation and Extension substantial
reflection
- Though excluded in this investigation, quadratic, cubic and sine models can be tested to
predict GDP more accurately in the short run.
- For simplicity of calculations, the maximum value M in the logistic model was taken
from secondary sources. However, the extent to which this affects the accuracy of
predictions cannot be determined.

18
 - A variety of more complex models such as the least squares regression could be
attempted to extend the investigation.
- Modeling the GDP growth rate of India, instead of GDP itself would resolve the
reliability issue that arose with excluding 2020 and 2021 due to COVID-19, as those data
points can be included. It was an idea initially considered to be the topic of this
investigation.

Figure 7: India's GDP growth rate (GDP Growth (Annual %) - India | Data)

However, upon realizing that the scope was far too complex due to the distribution of the
data points, India’s GDP instead of GDP growth rate was selected, but it could be
explored during further studies in mathematics.

VII) Bibliography

4.79+5.79xe^(−2019y)=x(E^(−1990y)) - Wolfram|Alpha. https://www.wolframalpha.com.

Accessed 16 Mar. 2022.

Dellink, Rob, et al. Consequences of Climate Change Damages for Economic Growth: A
Dynamic Quantitative Assessment. OECD, 30 June 2014. OECD iLibrary,
https://doi.org/10.1787/5jz2bxb8kmf3-en. Accessed 14 Mar. 2022.

“Desmos | Graphing Calculator.” Desmos, https://www.desmos.com/calculator. Accessed 17 Jan.

2022.

19
“Domestic Product - GDP Long-Term Forecast - OECD Data.” TheOECD,
http://data.oecd.org/gdp/gdp-long-term-forecast.htm. Accessed 14 Mar. 2022.

GDP (Constant 2015 US$) - India | Data.

https://data.worldbank.org/indicator/NY.GDP.MKTP.KD?locations=IN. Accessed 10 Jan. 2022.

GDP Growth (Annual %) - India | Data.

https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?locations=IN. Accessed 20 Mar.
2022.

“Is Infinite Economic Growth Possible on a Finite Planet?” LSE Business Review, 9 Nov. 2021,
https://blogs.lse.ac.uk/businessreview/2021/11/09/is-infinite-economic-growth-possible-on-a-
finite-planet/. Accessed 21 Jan. 2022.

OECD. Purchasing Power Parities (PPP). Organisation for Economic Co-operation and
Development, 2017. OECD iLibrary, https://www.oecd-ilibrary.org/finance-and-
investment/purchasing-power-parities-ppp/indicator/english_1290ee5a-en. Accessed 14 Mar.
2022.

VIII) Appendices
1. Appendix A - Chi-square calculations for the Exponential Model

Year Actual GDP Estimated GDP o-e (o-e)^2 (o-e)^2/e

/Billion units /Billion units (e)
(o)
1960 136.368 75.54880665 60.81919335 3698.97428 48.96138595
1961 141.445 80.26863068 61.17636932 3742.548163 46.62528975
1962 145.591 85.28332024 60.30767976 3637.016238 42.64627864
1963 154.318 90.61129671 63.70670329 4058.544044 44.79070702
1964 165.819 96.27213233 69.54686767 4836.766802 50.24056999
1965 161.448 102.2866221 59.16137791 3500.068636 34.21824443
1966 161.359 108.6768601 52.68213988 2775.407863 25.53816755
1967 173.987 115.4663208 58.52067915 3424.669888 29.65947008
1968 179.882 122.6799453 57.20205474 3272.075066 26.67163781
1969 191.645 130.3442325 61.30076751 3757.784098 28.82969216
1970 201.529 138.4873372 63.04166283 3974.251253 28.69757867

20
1971 204.84 147.1391729 57.70082714 3329.385453 22.62745799
1972 203.707 156.331522 47.37547805 2244.43592 14.35689931
1973 210.42 166.0981524 44.32184761 1964.426176 11.82689962
1974 212.914 176.4749417 36.43905828 1327.804968 7.524042537
1975 232.395 187.5000089 44.89499107 2015.560223 10.74965401
1976 236.26 199.2138544 37.04614558 1372.416902 6.889163939
1977 253.4 211.6595088 41.74049118 1742.268604 8.231468616
1978 267.876 224.882691 42.99330897 1848.424616 8.219505946
1979 253.844 238.9319762 14.91202379 222.3684535 0.930676827
1980 270.943 253.8589742 17.08402582 291.8639384 1.149708965
1981 287.216 269.718519 17.49748098 306.1618406 1.135116127
1982 297.199 286.5688705 10.63012947 112.9996526 0.394319357
1983 318.861 304.4719282 14.38907181 207.0453876 0.680014702
1984 331.044 323.4934586 7.550541393 57.01067533 0.176234399
1985 348.438 343.7033371 4.734662935 22.4170331 0.065222041
1986 365.082 365.1758042 -0.09380423 0.008799234 2.40959E-05
1987 379.559 387.9897389 -8.430738864 71.0773578 0.183193911
1988 416.102 412.2289476 3.873052407 15.00053495 0.036388844
1989 440.849 437.9824728 2.866527239 8.216978414 0.018760975
1990 465.243 465.3449195 -0.10191953 0.010387591 2.23223E-05
1991 470.16 494.4168034 -24.25680341 588.3925117 1.190073856
1992 495.936 525.3049195 -29.3689195 862.5334328 1.641967171
1993 519.496 558.1227348 -38.62673481 1492.024642 2.673291283
1994 554.089 592.990805 -38.90180504 1513.350436 2.552063915
1995 596.059 630.0372175 -33.97821751 1154.519265 1.832462008
1996 641.058 669.3980616 -28.3400616 803.1590916 1.199822852
1997 667.02 711.2179288 -44.19792876 1953.456907 2.746636196
1998 708.271 755.6504436 -47.37944359 2244.811675 2.970701194
1999 770.923 802.8588282 -31.93582823 1019.897125 1.27033183
2000 800.534 853.0165019 -52.48250192 2754.413008 3.229026638
2001 839.152 906.3077181 -67.15571807 4509.89047 4.97611394

21
2002 871.073 962.9282411 -91.85524112 8437.385321 8.762216083
2003 939.543 1023.086066 -83.54306564 6979.443816 6.82195179
2004 1014 1087.00218 -73.00218044 5329.31835 4.902766936
2005 1094 1154.91138 -60.91138036 3710.196257 3.21253762
2006 1183 1227.063129 -44.06312873 1941.559314 1.582281521
2007 1273 1303.722474 -30.72247387 943.8704006 0.723981077
2008 1312 1385.171023 -73.17102264 5353.998554 3.865225641
2009 1416 1471.707975 -55.70797498 3103.378477 2.108691758
2010 1536 1563.651223 -27.65122301 764.5901341 0.488977416
2011 1616 1661.338519 -45.33851878 2055.581285 1.237304295
2012 1705 1765.128715 -60.12871499 3615.462366 2.048271231
2013 1813 1875.403083 -62.40308328 3894.144802 2.076430842
2014 1948 1992.566715 -44.56671477 1986.192065 0.996800785
2015 2104 2117.050008 -13.05000818 170.3027135 0.080443406
2016 2277 2249.310251 27.68974909 766.7222049 0.34086992
2017 2432 2389.833299 42.16670117 1778.030688 0.743997788
2018 2591 2539.135361 51.86463889 2689.940767 1.059392425
2019 2696 2697.764896 -1.764896487 3.114859611 0.001154608
sum - 569.4095826

2. Appendix B - Chi-square calculations for the Logistic Model

Year Actual GDP Estimated GDP o-e (o-e)^2 (o-e)^2/e

/Billion units (o) /Billion units (e)
1960 136.368 74.07984018 62.28815982 3879.814854 52.37342365
1961 141.445 78.81144473 62.63355527 3922.962245 49.7765554
1962 145.591 83.84480857 61.74619143 3812.592157 45.47201218
1963 154.318 89.19911631 65.11888369 4240.469013 47.53936124
1964 165.819 94.89476437 70.92423563 5030.2472 53.00869056
1965 161.448 100.9534364 60.49456362 3659.592228 36.25029875
1966 161.359 107.3981833 53.9608167 2911.769739 27.11190869

22
1967 173.987 114.2535083 59.73349172 3568.090033 31.22958836
1968 179.882 121.5454566 58.33654336 3403.152291 27.99900864
1969 191.645 129.3017112 62.34328879 3886.685657 30.05904269
1970 201.529 137.5516934 63.97730665 4093.095766 29.75678209
1971 204.84 146.32667 58.51333004 3423.809793 23.3983989
1972 203.707 155.6598667 48.04713327 2308.527015 14.83058584
1973 210.42 165.5865881 44.83341187 2010.03482 12.13887455
1974 212.914 176.1443442 36.7696558 1352.007587 7.675566272
1975 232.395 187.3729849 45.02201514 2026.981848 10.81789805
1976 236.26 199.3148417 36.94515827 1364.94472 6.848184048
1977 253.4 212.0148782 41.38512183 1712.728309 8.078340177
1978 267.876 225.5208477 42.35515229 1793.958926 7.954736532
1979 253.844 239.8834613 13.96053873 194.8966416 0.812463854
1980 270.943 255.1565638 15.78643619 249.2115677 0.976700595
1981 287.216 271.3973204 15.81867957 250.2306232 0.922008452
1982 297.199 288.6664127 8.532587261 72.80504537 0.252211695
1983 318.861 307.0282455 11.83275446 140.0140782 0.456029959
1984 331.044 326.5511645 4.492835475 20.18557061 0.061814419
1985 348.438 347.3076852 1.13031478 1.277611501 0.003678616
1986 365.082 369.3747336 -4.292733597 18.42756173 0.049888528
1987 379.559 392.8338988 -13.27489876 176.2229372 0.448594018
1988 416.102 417.7716981 -1.669698069 2.787891642 0.006673242
1989 440.849 444.2798549 -3.430854923 11.7707655 0.026494034
1990 465.243 472.4555897 -7.212589665 52.02144968 0.110108655
1991 470.16 502.4019237 -32.24192369 1039.541643 2.069143437
1992 495.936 534.227997 -38.29199702 1466.277036 2.744665281
1993 519.496 568.0493995 -48.5533995 2357.432603 4.150048579
1994 554.089 603.9885156 -49.89951559 2489.961656 4.122531459
1995 596.059 642.1748827 -46.11588273 2126.67464 3.31167521
1996 641.058 682.7455631 -41.68756315 1737.852921 2.545388817
1997 667.02 725.8455288 -58.82552884 3460.442843 4.767464572

23
1998 708.271 771.6280592 -63.35705917 4014.116947 5.202139683
1999 770.923 820.2551507 -49.33215066 2433.661089 2.966956181
2000 800.534 871.8979381 -71.36393806 5092.811656 5.841063998
2001 839.152 926.7371257 -87.58512567 7671.154239 8.277594613
2002 871.073 984.9634278 -113.8904278 12971.02954 13.16904686
2003 939.543 1046.778016 -107.2350164 11499.34874 10.98547023
2004 1014 1112.392975 -98.39297473 9681.177477 8.703019254
2005 1094 1182.031754 -88.03175406 7749.589723 6.556160354
2006 1183 1255.929631 -72.929631 5318.731078 4.234895767
2007 1273 1334.334163 -61.33416273 3761.879518 2.819293415
2008 1312 1417.505636 -105.5056361 11131.43925 7.852835971
2009 1416 1505.717507 -89.71750668 8049.231005 5.345777657
2010 1536 1599.256823 -63.25682308 4001.425667 2.502053209
2011 1616 1698.424631 -82.42463101 6793.819797 4.000071403
2012 1705 1803.536351 -98.53635117 9709.412502 5.383541339
2013 1813 1914.922124 -101.9221241 10388.11938 5.424826029
2014 1948 2032.927114 -84.92711431 7212.614746 3.547896378
2015 2104 2157.911765 -53.91176517 2906.478423 1.346894007
2016 2277 2290.251995 -13.25199523 175.6153775 0.0766795
2017 2432 2430.339326 1.660674467 2.757839685 0.001134755
2018 2591 2578.580927 12.41907345 154.2333854 0.059813281
2019 2696 2735.399572 -39.39957236 1552.326302 0.567495264
Sum - 655.0214992

24